All docs pages (49)
Smart Data Generator
Fill an empty sandbox with realistic, schema-aware test data: relationships and picklists respected, active validation rules flagged.
Updated 2026-06-08
A fresh developer sandbox has no data, and building 50 related records by hand to test a flow is an afternoon you will not get back. Smart Data Generator creates realistic test data that respects your schema, so you can populate an org in minutes instead of hours.
Zero configuration
Point it at an object and it reads the object's describe metadata to learn the rules before generating anything:
- Required fields are filled. A required lookup needs its parent object in the run, or existing parent records in the org; otherwise preview and insert are blocked and a warning names the missing lookup. Salesforce can still reject a record for other reasons, and those show in the error log.
- Relationships are resolved in order — parents are created before the children that reference them.
- Active validation rules are listed as warnings for each object, so you can adjust values before you generate. Generated values are not changed to satisfy them.
- Picklists (including dependent picklists) draw from the actual allowed values.
Realistic values
Generated data looks like real data, not Test 1, Test 2. Names, addresses, dates, and currency values are believable, and naming is object-aware:
Account.Namebecomes a company name.Opportunity.Namebecomes something like "Acme - Q4 Renewal".Product2.Namebecomes a product name.Case.Subjectbecomes a realistic case subject.Task.Subjectbecomes a realistic task line.
Custom objects fall back to sensible generic names.
Wave-by-wave insertion
Records insert in waves using sObject Collections at 200 records per batch, in dependency order, so a parent object and its children land in the right sequence. A live execution panel shows progress per wave, and you can cancel mid-run. Cancel stops before the next 200-record call; a call already sent still finishes, and its records stay in the rollback ledger.
Safe by default
Smart Data Generator is built not to waste your API quota or corrupt data:
- It halts on fatal errors — an expired session, a request-limit breach, or a maintenance window stops the run instead of retrying into the ceiling.
- Transient failures retry with backoff; permanent ones surface row-by-row in the error log.
- A rollback ledger records everything inserted, so a generation run can be undone — even a run that halted partway.
Typical workflow
- Select the object (and any related objects) you want to populate.
- Choose how many records per object.
- Review the API-call estimate.
- Generate — and watch the waves insert.
- Keep the data, or roll the whole run back from the ledger.
When a link isn't set
Lookups between generated records (for example Parent Account in a hierarchy) are set in a second pass after the insert, because a record can only point at its parent once the parent exists. If Salesforce rejects one of those links, for example a validation rule, the record stays inserted and only the link is left blank. The Needs linking tab lists each record. From version 1.1.13 it also shows Salesforce's reason for each one; before that, Copy errors copies the reasons to your clipboard. If the link points at a record that failed to insert, fix and retry that record and the link is set automatically.
Walkthrough
Moving and making test data follows a real run step by step: 5 accounts and 10 contacts generated in a Consumer Goods org, the 3 links a validation rule refused, and Undo all.
Related reading
- Object Data Dictionary (in the extension's Metadata menu) — inspect the fields of the object you are generating.
- SOQL Builder — query and verify the generated records.
